How about a book on Delphi (or insert your favorite language here)

I have a long history with Delphi. I remember being in my friend Zack Urlocker's office at Borland when he showed it to me for the first time. I told him that I liked the name, and that he should keep it (it was originally a code name for the project; I suspect many others told him they liked the name as well, but I like to think I had some influence...). Every once in awhile I would take a dive into it, and create the odd program. The summer of 2000, after the Borland conference, I sponsored Marco Cantu to give an advanced Delphi seminar. The best equivalent for a "Thinking in Delphi" is Marco's collection of books.

Even though Pascal, and Turbo Pascal, were my first languages, I have had so much experience in C-ish languages that it's awfully difficult for me to go back to Pascal syntax. So if I did need to create a Windows-specific program and I couldn't use Python (my favorite language, which I use all the time) then I would probably use C++ Builder, since I am much more fluent in that syntax.

And books, well. They're getting harder to do, and taking longer, so I get very careful about what I think I can do. I suspect it has to do with what I want from a book, which is based on all the feedback I get about the books I put up on the Web. I'm sure my process could be improved but in general I'm getting slower rather than faster at producing books (and of course I'm slowed down by all the other things I do - but somehow Martin Fowler is able to pull it off...).

As far as a book on a specific language, I carefully consider the languages that are out there and which ones seem likely to become interesting enough to justify a book. Although there are many languages that have contributed to computing, many of them have had their shot at glory and didn't take. Typically, this means there was a big marketing push at some point, a lot of buzz for awhile (sometimes quite awhile, as in the case of SmallTalk), and then the language faded from the scene for one reason or another. Examples of this in the OO realm are Eiffel and Objective-C, both of which tried to be commercial languages and are, long after the fact, now available in free versions, but too late to really do anyone any good. (Python is different -- it's always been grassroots and has never had any big marketing push behind it, so it's more like Linux. Slow and steady increments). Yes, Squeak is the most interesting implementation of Smalltalk and there's a book or two out about it, but I still don't see Smalltalk making a big recovery. Ruby still looks to me like a language that was invented because you couldn't (rationally) use objects or references in Perl, so it has Perl-ish syntax, which I don't think is very helpful (someday I may learn it, but Python continues to be far more compelling to me). Long ago I used Lisp, but probably won't again. So no, I'm probably not going to write books on any of those languages except for Python. One of the important things about being a programmer is learning to make good use of your time, and choosing the right languages is an essential aspect of that.